Investigation of Core Structure and Stability of Human Pyruvate Dehydrogenase Complex: A Coarse-Grained Approach.

The human pyruvate dehydrogenase complex (hPDC) is a large macromolecular machine, and its unique structural and functional properties make it a versatile target for manipulation aiming for the design of new types of artificial multienzyme cascades. However, model-based and hence systematic understanding of the structure-function relationship of this kind of complexes is yet poor. However, with new structure data, modeling techniques, and increasing computation power available, this shortfall is about to cease. Recently, we have built new atomistic models of E2/E3BP, the two subunits of the massive hPDC core. Here, we present developed coarse-grained models of the same, on the basis of which we built and simulated the full core of hPDC, which is so far the first simulation of the catalytic core of any member in the branched-chain α-keto acid dehydrogenase complex family. We explored the stability of two previously proposed substitutional models of hPDC core: 40E2+20E3BP and 48E2+12E3BP. Our molecular dynamics simulations showed a higher stability and sphericity for the second model. Our core's radius of gyration was found to be in strong agreement with previously published experimental dynamic light scattering (DLS) data. Finally, in the direction of our experimental effort to design active minimized complexes, we simulated C-terminal truncated E2/E3BP cores of different lengths, which clearly showed the instability of the core assembly and symmetry due to subunit separations. This correlated very well with the findings of our experimental investigations by analysis of DLS data for variable truncation lengths. The use of polarizable water and an increased total ion concentration did not lead to a substantially different initial stability of the truncated mutants compared to that of standard MARTINI water; however, a different rearrangement behavior of the fragments was observed. The obtained structure models will serve as a basis for full complex simulations in the future, providing the possibility for the model-assisted targeted manipulation of such a complex enzymatic system.